| The high-water-level coal-grain complex area is endowed with abundant coal resources and a large amount of cultivated land resources.Under the premise of ensuring that the quality and quantity of cultivated land are not damaged as much as possible,the mining of coal resources will inevitably have an impact on the environment of the mining area.The buried depth of the underground water in the high-water-level mining area is shallow,and underground mining will cause the surface to sink,waterlogging will be generated in severe cases.This phenomenon will cause extensive damage to cultivated land,make crop production reduction or crop failure,and aggravate the contradiction between human and land.Improving crop yield in high-water-level mining areas is the main purpose of exploring the damage to cultivated land in mining areas by underground mining.It is of great significance to study the changing mechanism of crop yield affected by underground mining,which can not only provide reference for the formulation of production plans of coal mining enterprises,but also provide reference for changes in field management measures,in order to minimize the impact of underground mining on crop yield.Biomass can well characterize the growth status of crops.Crop biomass at the maturity stage of crops can generally be regarded as crop yield.Crop biomass is closely related to crop yield,and crop biomass is more convenient to track changes in crop growth conditions in different stage.In this study,the surface cultivated land affected by the underground mining of a certain working face is the research object,and the period from the mining start of the working face to the stable state is the period,and the purpose of this study is to construct a dynamic monitoring model of cultivated land damage based on the integration of underground-surface-aboveground factors.In this thesis,the UAV multispectral and RGB images are used as remote sensing data sources,combined with the field measured data,firstly the impact of underground mining on surface elevation was studied;secondly the impact of underground mining on soil moisture was studied;and then the impact of underground mining on crop biomass was studied;finally,a dynamic monitoring model of cultivated land damage combined with surface elevation,soil moisture and underground mining parameters was established to realize real-time monitoring of cultivated land crops under the influence of underground mining.The main research contents and conclusions of the thesis are as follows:(1)Surface elevation extraction is the basis for the changes of other indicators in the study area.According to the vegetation coverage of crops in different periods,a surface elevation extraction method is proposed,that is,when the vegetation coverage is high,the combination of automatic extraction and manual culling is used to extract the surface elevation,and when the vegetation coverage is medium,the fuzzy C-means clustering method is used to extract the surface elevation.The results show that:(1)In the case of high vegetation coverage,the accuracy of the extraction results of the surface elevation is lower than that of the extraction results in other cases,indicating that the extraction results of the surface elevation are affected by the vegetation coverage.(2)The accuracy R~2of the extraction results of the fuzzy C-means clustering method reaches 0.97,indicating that this method is suitable for the extraction of surface elevation when the vegetation coverage is moderate.(3)With the advancement of underground mining,the position of the subsidence center in the study area gradually moved to the right from the direction of the opening of the working face,and the influence range of the subsidence center gradually expanded.When the working face is advanced to 610m,the surface elevation changes significantly,and the position and scope of the subsidence center are basically stable about 60 days after the end of mining(2)An Elevation-Vegetation coverage-Soil moisture Drought Index(EVMDI)was constructed to monitor soil moisture changes in the study area.Using Elevation-Vegetation coverage-Soil Moisture Drought Index(EVMDI),Vegetation Condition Index(VCI),Perpendicular Drought Index(PDI),Modified Perpendicular Drought Index(MPDI),Soil Moisture Monitoring of Remote Sensing(SMMRS),The Proposed Ratio Dryness Monitoring Index(RDMI)to estimate soil moisture,the monitoring results were compared and analyzed.Finally,the temporal and spatial changes of soil moisture were analyzed.The results show that:(1)The estimation accuracy of different drought indices for estimating drought conditions in the study area is quite different,and the estimation accuracy of EVMDI drought index is better than that of other drought indices.(2)The monitoring results of the drought index under different crop coverage are different.Whether the surface crops coverage is maize or wheat,EVMDI has the best drought status estimation results,and the absolute value of R is greater than 0.57,indicating that the EVMDI is robust.The RDMI drought index is better for the estimation result of the drought status when the land cover is wheat,and the VCI index is better for the estimation result of the drought status when the land cover is maize.(3)At the beginning of underground mining,the drought degree of most areas in the study area were normal,and a few area were humidity.With the advancement of underground mining,the degree of drought in the edge area of the study area increased,and the degree of humidity in the central area increased.At the same time,the increased area of humidity was larger than that of the increased area with drought,indicating that the closer to the mining subsidence center,the greater the degree of humidity.(4)Soil moisture changes nonlinearly under the influence of underground mining.(3)Biomass inversion model combining spectral information,texture features and plant height were constructed.First,the variance inflation factor and the Pearson correlation coefficient were used to optimize the vegetation index and texture features,and then combined with the plant height,multiple linear regression,partial least squares regression,random forest regression and Gaussian process regression were used to invert the crop biomass,finally,the changes of the spatial characteristics of biomass in the study area and the impact on underground mining were analyzed.The results are as follows:(1)Multiple linear regression was the optimal inversion model at the jointing stage of maize,and random forest regression was the optimal regression model at the heading stage and grain filling stage of wheat,and the coefficients of determination R~2of the three was greater than 0.86.(2)The biomass at the jointing stage of maize was not affected,the proportion of the area with slight biomass damage at the heading stage of wheat was accounted for 78.19%,and the proportion of the area with normal damage was accounted for 17.33%.During the wheat grain filling period,the proportion of the area with slight biomass damage was accounted for 23.65%,the proportion of the area with normal damage was accounted for 41.59%,and the proportion of the area with severe damage was accounted for 27.65%.(3)From the subsidence center to the edge area,the biomass change trend is to increase first,then decrease,and then increase,showing a nonlinear change,and the impact of surface subsidence on biomass is consistent with that from the subsidence center to the edge area.(4)The impact of biomass on the surface elevation changes linearly,and the surface elevation in the subsidence center is the lowest and the biomass is the smallest.(4)Dynamic monitoring model for farmland damage was constructed.First,the predicted value of dynamic subsidence at any point on the surface was calculated,and then the dynamic response model of surface elevation and soil moisture was established..The results are as follows:(1)The mean median error of the surface elevation dynamic response model is±91.25mm and the mean relative error is 4.56%.(2)The correlation coefficients between the predicted value of the EVMDI drought index and the measured soil moisture are all greater than 0.60,and the significance P is less than 0.01.(3)The impact of surface elevation on cultivated land damage was significantly greater than that of soil moisture.(4)The dynamic monitoring model of cultivated land damage can better predict the damage degree of aboveground crops corresponding to different advancing distances of underground mining,and the model determination coefficient R~2is 0.68. |
PDF Full Text Request